We are pleased to introduce Emma Holliday, a food science master’s degree student in the Food Science and Human Nutrition program at The University of Florida. She is our latest guest in the FSHN Research Journeys series, which follows graduate students’ research. Emma is keen to ensure the food we eat is safe and that we have quick, effective ways to tell if food may make us ill. In her research, she is studying the development of an AI-based pathogen sensor that could give food suppliers the power to easily determine if food is safe for consumers to enjoy.

Emma: Have you ever found old lettuce in your refrigerator and struggled to know if it was safe to eat? The real struggle lies in the uncertainty. Without knowing if the lettuce is safe you can either try your luck and eat it, risking illness from foodborne pathogens, or throw the lettuce away, potentially wasting it. It would be great if there was a fast and convenient way to gain insight into the safety of your lettuce.

Informed decision-making saves you stress, time, and money. In the food supply chain, the need for rapid detection of foodborne pathogens is amplified. The large scale of the global food supply results in increased damage from making the wrong decision, including widespread foodborne illness or costly food waste. Sensors have great potential in providing food suppliers the answers they need to make smart food safety decisions.

The significance of foodborne illness and pathogen detection

Foodborne illnesses have detrimental health, social, and economic consequences. They can cause painful, uncomfortable symptoms and even death. For example, Salmonella infection often results in diarrhea, fever, and stomach cramps.¹ The Centers for Disease Control estimates that foodborne pathogens yearly result in 48 million illnesses and 3000 deaths.² Foodborne illnesses also reduce the public’s trust in food production systems. In addition, foodborne illness outbreaks are expensive, costing $15.6 billion each year in the United States for medical expenses and lost productivity.^3,4 Salmonella has the highest cost of foodborne bacteria each year, with 1.35 million cases and a financial cost of at least $3.3 billion.^4,5

Preventing foodborne pathogens in foods is essential, but they can be difficult to control due to the complex and dynamic conditions throughout the food supply chain and the adaptability of bacteria. Currently, there are few controls over food during distribution. This deficiency creates vulnerabilities in the food supply chain. Therefore, we must have measures during distribution to monitor the food and ensure that the final product is safe for consumers.

The shortcomings of other foodborne pathogen sensors

Pathogen sensors facilitate foodborne pathogen detection. Many sensors directly contact food samples, destroying that food sample. On the other hand, non-destructive sensors enable continuous monitoring of foodborne pathogens throughout the supply chain, such as in smart packaging. Sensors that detect foodborne pathogens non-destructively have been developed by targeting gases emitted from bacteria, but they currently contain toxic chemicals that cannot be used near food.^6,7,8,9

The food industry needs a sensor that can overcome these struggles. I plan to use non-toxic elements and artificial intelligence (AI) to create a sensor that addresses the shortcomings of other sensors. This sensor will detect pathogens quickly and continuously and contain only non-toxic elements to maximize its potential in food safety monitoring and smart food packaging.

My journey to foodborne pathogen sensors

I first learned about food safety during high school. I studied restaurant food safety principles in my culinary arts classes and appreciated their application in my home kitchen. When I took microbiology in college, I came to class excited to learn more. Food microbiology research has allowed me to combine these two interests. More specifically, I became interested in foodborne pathogen detection systems through an undergraduate research opportunity at UF under Assistant professor Dr. Boce Zhang and Postdoctoral Research Associate Dr. Zhen Jia. While working in their lab, I aided with a project developing a sensor for detecting low concentrations of Salmonella in peanut butter.

I wanted to continue this work in my graduate studies under Dr. Zhang because I find it practical, important, and challenging. I enjoy doing research with tangible results and direct food industry applications. Overall, I want to use my degree to solve widespread problems in the food industry, and this project allows me to chip away at one major issue.

Developing an AI-based sensor that can detect Salmonella

My project will be the development of an AI-based sensor that can detect and quantify Salmonella in food. I will use a non-toxic colorimetric sensing system, paper microfluidics, and machine learning AI to sense the gases emitted by Salmonella in a food sample. My sensor relies on the reactions of an array of non-toxic dyes with gases emitted by Salmonella which cause color changes on a piece of paper. The resulting sensor will hopefully reduce foodborne illness by detecting Salmonella in food products before reaching consumers. This early detection gives food companies more control over the microbial safety of their products.

I will first create the components of the sensor, which will consist of paper with spots of non-toxic dyes. This framework is based on previous successful sensors.^6,7 Paper-based sensors are common in point-of-care medical diagnostic devices because of their portability, relatively simple use, and low cost, making them appealing in the food industry.^10,11 Because one primary novelty of the sensor lies in the non-toxic dyes, dye selection will be a major part of the research. The dyes will be screened beforehand to test their reactivity with the gases and ensure that several non-toxic dyes show color changes.

Testing the AI-based sensor on Salmonella

After sensor construction, I will expose the device to Salmonella samples, allowing interaction between the bacterial gases and the dye. Ideally, we will detect a measurable color change on the paper. The results of this trial will be used to develop the AI model, namely a neural network, that can process the complex color data from an array of dye spots to accurately detect and quantify the bacteria.

In the next step, I will test the device on food samples contaminated with Salmonella. I will need to address the increase in the samples’ complexity as background bacteria in the food samples can affect the sensor’s ability to detect and quantify the pathogen. Finally, I will optimize the sensor by adjusting the sensor design or the neural network algorithm. Adjustments may include changing the concentrations of dyes and tweaking the code for the neural network.

Overcoming challenges of sensor development

Microbiology work, including the lab work required for this project, can be highly sensitive and precise. Many factors, such as hand coordination and pipetting variability, may contribute to contamination or other errors. With more laboratory practice, training, and advice from my lab mates, I hope to improve my technique, increase my accuracy, and effectively troubleshoot. For example, I will be learning molecular techniques, including DNA sequencing, from my lab mate, Tingting Gu. This technique requires precise hand movement involving minute volumes of liquid.

I may also experience challenges with the artificial intelligence aspect of my project because I lack a strong computer science background. To overcome these challenges, I plan to include coding and machine learning in my coursework to overcome this challenge. In addition, I will collaborate with experts in computer science to develop the code.

This food safety research has practical applications now but ultimately looks toward the future. At its core, it aims to solve the large practical issue of foodborne illness from commercially produced foods. It also aims to contribute to the growth of novel interdisciplinary technology by expanding the application of AI in food science. We can all hope for a future in which we will not have to choose between risking our health and being wasteful every time we see questionable lettuce.

Emma Holliday is a first-year master’s degree student from Orlando, FL. She received her undergraduate degree in food science from the University of Florida, where she became passionate about food microbiology. She hopes to spend her time in graduate school doing high-impact research, gaining technical skills, and growing professionally. In her free time, she enjoys music, cooking, comedy, and being outdoors. She wants to thank her advisor Dr. Zhang and mentor Dr. Jia for believing in her and helping her invest in her future.

References

1. Symptoms | Salmonella | CDC. (2019, December 12). https://www.cdc.gov/salmonella/general/salmonella-symptoms.html
2. Burden of Foodborne Illness: Overview | Estimates of Foodborne Illness | CDC. (2018, November 15). https://www.cdc.gov/foodborneburden/estimates-overview.html
3. CDC and Food Safety | CDC. (2023, March 15). https://www.cdc.gov/foodsafety/cdc-and-food-safety.html
4. Hoffmann, S., Batz, M. B., & Morris, J. G. (2012). Annual Cost of Illness and Quality-Adjusted Life Year Losses in the United States Due to 14 Foodborne Pathogens. Journal of Food Protection, 75(7), 1292–1302. https://doi.org/10.4315/0362-028X.JFP-11-417
5. Medicine, C. for V. (2023). Get the Facts about Salmonella. FDA. https://www.fda.gov/animal-veterinary/animal-health-literacy/get-facts-about-salmonella
6. Jia, Z., Luo, Y., Wang, D., Dinh, Q. N., Lin, S., Sharma, A., Block, E. M., Yang, M., Gu, T., Pearlstein, A. J., Yu, H., & Zhang, B. (2021). Nondestructive multiplex detection of foodborne pathogens with background microflora and symbiosis using a paper chromogenic array and advanced neural network. Biosensors and Bioelectronics, 183, 113209. https://doi.org/10.1016/j.bios.2021.113209
7. Jia, Z., Luo, Y., Wang, D., Holliday, E., Sharma, A., Green, M. M., Roche, M. R., Thompson-Witrick, K., Flock, G., Pearlstein, A. J., Yu, H., & Zhang, B. (2024). Surveillance of pathogenic bacteria on a food matrix using machine-learning-enabled paper chromogenic arrays. Biosensors and Bioelectronics, 248, 115999. https://doi.org/10.1016/j.bios.2024.115999
8. Li, H., Geng, W., Zhang, M., He, Z., Haruna, S. A., Ouyang, Q., & Chen, Q. (2022). Qualitative and quantitative analysis of volatile metabolites of foodborne pathogens using colorimetric-bionic sensor coupled robust models. Microchemical Journal, 177, 107282. https://doi.org/10.1016/j.microc.2022.107282
9. Yang, M., Liu, X., Luo, Y., Pearlstein, A. J., Wang, S., Dillow, H., Reed, K., Jia, Z., Sharma, A., Zhou, B., Pearlstein, D., Yu, H., & Zhang, B. (2021). Machine learning-enabled non-destructive paper chromogenic array detection of multiplexed viable pathogens on food. Nature Food, 2(2), 110–117. https://doi.org/10.1038/s43016-021-00229-5
10. Nishat, S., Jafry, A. T., Martinez, A. W., & Awan, F. R. (2021). Paper-based microfluidics: Simplified fabrication and assay methods. Sensors and Actuators B: Chemical, 336, 129681. https://doi.org/10.1016/j.snb.2021.129681
11. Yao, Z., Coatsworth, P., Shi, X., Zhi, J., Hu, L., Yan, R., Güder, F., & Yu, H.-D. (2022). Paper-based sensors for diagnostics, human activity monitoring, food safety and environmental detection. Sensors & Diagnostics, 1(3), 312–342. https://doi.org/10.1039/D2SD00017B

Looking for more posts exploring graduate research projects in the FSHN Department at the University of Florida?

Dive into the Research Journeys of other graduate students below.

M.S. Food Science
M.S. Nutritional Sciences
Ph.D. Food Science
Ph.D. Nutritional Sciences

1